A Full-Stack MERN-Based Real-Time Language Exchange Platform — Design, Challenges and Implementation

Abstract

Language learning through real human conversation remains one of the most effective yet underserved methods of achieving fluency. Students and language enthusiasts worldwide face significant challenges in finding suitable native speaking partners outside formal educational settings. Existing platforms either focus on structured lesson delivery without human inter- action or provide fragmented communication tools that require users to switch between multiple applications during a single practice session. Traditional language learning services rely heavily on classroom instruction, expensive private tutoring, and static web content, which become inefficient and inaccessible for a large population of self-directed learners.
This paper presents the design, development, and implemen- tation of Serenly, a full-stack web application built on the MERN stack — MongoDB, Express.js, React, and Node.js — designed to connect language learners with native speaking partners through a unified platform offering real-time one-on-one chat and video calling. The proposed platform provides secure user authentication, a structured onboarding flow capturing linguistic profiles, a language-aware partner discovery mechanism, a friend request and acceptance system, real-time one-on-one messaging powered by the Stream Chat SDK, and one-on-one video calling through the Stream Video SDK, all within a single cohesive web application.
The platform is developed using React 19, Vite 6, TanStack Query 5, Tailwind CSS, DaisyUI, Zustand, Node.js, Express 4, MongoDB with Mongoose 8, and JWT-based authentication with HTTP-only cookies. The system supports 32 switchable UI themes with persisted preferences, responsive layouts across all device sizes, and secure cross-origin deployment with the frontend on Vercel and the backend on Render. Experimental analysis demonstrates that the platform improves language learning accessibility, reduces dependency on expensive tutoring services, enables real-time human conversation practice, and simplifies the process of finding and connecting with suitable language partners